Search

Vibration-based fault diagnostics combined with deep learning approaches has promising applications in detecting and diagnosing faults in wind turbine gearboxes. Specifically when time series vibration data is transformed to a 2-dimensional cyclic spectral coherence maps, the accuracy of deep neural networks in classifying faults increases. Nevertheless, standard deep learning techniques are vulnerable to inaccurate predictions when tested with new data originating from unseen faults or unusual operating conditions. To address some of these shortcomings in the context of wind turbine gearboxes, this paper investigates fault diagnostics using Bayesian convolutional neural network which provide accurate results with uncertainty bounds reducing wrong overconfident classifications. The performance of Bayesian and standard neural networks is compared using a simulation-based dataset of acceleration signals generated from a multibody dynamic model of a 5 MW wind turbine. The framework proposed in this paper has relevance to fault detection and diagnosis in other rotating machinery applications. [ABSTRACT FROM AUTHOR]

To reduce wind turbine operations and maintenance costs, we present a machine learning framework for early damage detection in gearboxes based on the cyclostationary and kurtogram analysis of sensor data. The application focus is fault diagnostics in gearboxes under varying load conditions, particularly turbulent wind. Faults in the gearbox rotating components can leave their signatures in vibrations signals measured by accelerometers. We analyze data stemming from a simulated vibration response of a 5 MW multibody wind turbine model in a healthy and damaged scenarios and under different wind conditions. With cyclostationary and kurtogram analysis applied on acquired sensor data, we generate two types of 2D maps that highlight signatures related to the fault damage. Using these maps, convolutional neural networks are trained to identify faults, including those of small magnitude, in test data with a high accuracy. Benchmark test cases inspired by an NREL study are tested and faults successfully detected. [ABSTRACT FROM AUTHOR]

The objective of this paper is twofold: first to illustrate that nonlinear modal interactions, namely, a two-to-one internal resonance energy pump, can be exploited to improve the steady-state bandwidth of vibratory energy harvesters; and, second, to investigate the influence of key system's parameters on the steady-state bandwidth in the presence of the internal resonance. To achieve this objective, an L-shaped piezoelectric cantilevered harvester augmented with frequency tuning magnets is considered. The distance between the magnets is adjusted such that the second modal frequency of the structure is nearly twice its first modal frequency. This facilitates a nonlinear energy exchange between these two commensurate modes resulting in large-amplitude responses over a wider range of frequencies. The harvester is then subjected to a harmonic excitation with a frequency close to the first modal frequency, and the voltage-frequency response curves are generated. Results clearly illustrate an improved bandwidth and output voltage over a case which does not involve an internal resonance. A nonlinear model of the harvester is developed and validated against experimental findings. An approximate analytical solution of the model is obtained using perturbation methods and utilized to draw several conclusions regarding the influence of key design parameters on the harvester's bandwidth. [ABSTRACT FROM AUTHOR]

Depending on the shape of their potential energy function, nonlinear vibratory energy harvesters (VEHs) can be classified either as mono- or bi-stable. The complex dynamic responses of such harvesters make it difficult to accurately analyze their performance relative to one another. In this paper, we aim to tackle this issue by comparing the performance of these two classes in response to harmonic fixed-frequency excitations under optimal electric loading conditions. To achieve this goal, the mathematical model of a generic nonlinear VEH capable of operating in both of the mono- and bi-stable configurations is considered. Using perturbation methods, analytical expressions for the steady-state output power are obtained. The output power is then optimized with respect to the time constant ratio which represents a direct measure of the electric load. We study the variation of the optimal output power with the excitation frequency in both configurations for different potential shapes and excitation levels. It is shown that, tuning the optimal time constant ratio close to the excitation frequency results in maximum power transduction in both configurations. Results also illustrate that bi-stable VEHs produce higher power levels than mono-stable ones under optimal electric loading conditions regardless of the shape of the potential function. [ABSTRACT FROM AUTHOR]